The molecular impact of cigarette smoking resembles aging across tissues

Abstract

Background: Tobacco smoke is the main cause of preventable mortality worldwide. Smoking increases the risk of developing many diseases and has been proposed as an aging accelerator. Yet, the molecular mechanisms driving smoking-related health decline and aging acceleration in most tissues remain unexplored.

Methods: Here, we use data from the Genotype-Tissue Expression Project (GTEx) to perform a characterization of the effect of cigarette smoking across human tissues. We perform a multi-tissue analysis across 46 human tissues. Our multi-omics characterization includes analysis of gene expression, alternative splicing, DNA methylation, and histological alterations. We further analyze ex-smoker samples to assess the reversibility of these molecular alterations upon smoking cessation.

Results: We show that smoking impacts tissue architecture and triggers systemic inflammation. We find that in many tissues, the effects of smoking significantly overlap those of aging. Specifically, both age and smoking upregulate inflammatory genes and drive hypomethylation at enhancers (odds ratio (OR) = 2). In addition, we observe widespread smoking-driven hypermethylation at target regions of the Polycomb repressive complex (OR = 2), which is a well-known aging effect. Smoking-induced epigenetic changes overlap causal aging CpGs, suggesting that these methylation changes may directly mediate the aging acceleration observed in smokers. Finally, we find that smoking effects that are shared with aging are more persistent over time.

Conclusion: Overall, our multi-tissue and multi-omic analysis of the effects of cigarette smoking provides an extensive characterization of the impact of tobacco smoke across tissues and unravels the molecular mechanisms driving smoking-induced tissue homeostasis decline and aging acceleration.

Keywords: Aging; Cigarette smoking; DNA methylation; Gene expression; Histology images; Multi-omics; Multi-tissue.

Fig. 1

Individuals, tissues, and data modalities analyzed. For all tissues, RNA sequencing and histological images are available. Tissues highlighted in bold also include DNA methylation profiles. We classified GTEx individuals according to their smoking status

Fig. 2

Smoking differential gene expression and alternative splicing analysis. a Tissue sample size (left) and number of smoking-DEGs per tissue (right). b Number of smoking-DEGs per tissue at varying sample sizes while keeping an equal number of smokers and never smokers. c Number of tissues in which the smoking-DEGs are differentially expressed. Gene names are shown for smoking-DEGs in more than 9 tissues. d Number of tissue-specific and tissue-shared gene ontology enriched terms (FDR < 0.05, minimum gene count > 5). e Most enriched and summarized gene ontology terms (biological processes) for smoking up- (red) and down- (blue) regulated genes. f Tissue sample size (left) and number of smoking-DSEs per tissue (right). g PSI values in smokers and never-smokers for a skipped exon in CLDN7 (top) and visual representation of the two most expressed isoforms involved in the exon skipping event (bottom). Isoform 1 is ENST00000360325 and Isoform 2 is ENST00000538261

Fig. 3

Impact of smoking on histology. a Number of images analyzed per tissue and per donor smoking status. b Receiver operating characteristic (ROC) curves of the tissue histology classifiers. c Example images of a never smoker and a smoker’s lung. Brown macrophages are highlighted in red circles. d Estimated proportion of macrophages in never smokers, ex-smokers, and smokers. e Example images of a never smoker and smoker’s thyroid. Smoker thyroids have larger colloid-containing follicles. f Mean (per donor) of median (per tile) diameter of thyroid follicles. p-value obtained from a Wilcoxon test. g Median (per donor) of standard deviation (sd) (per tile) diameter of thyroid follicles. p-value obtained from a Wilcoxon test

Fig. 4

Impact of smoking and aging on gene expression. a Number of smoking-DEGs that are also DE with one demographic trait per tissue. Colored cells indicate higher-than-expected overlaps (Fisher’s exact tests; FDR < 0.05). Tissues are sorted from highest (top) to lowest (bottom) sample size. b Example of a gene with additive effects for smoking and age. c Tissues with significant concordance in the direction of change (up- or down-regulation) for smoking-age-DEGs across tissues (chi-squared tests; FDR < 0.05). Tissues are sorted from lowest (top) to highest (bottom) p-value. d Example of a gene with an interaction effect between smoking and age. e Expression of TDRD3 stratified by genotype at rs7924558 position and smoking status

Fig. 5

Association of smoking with DNA methylation. a Number of samples (left) and smoking-DMPs (right) across tissues. b Smoking-DMPs enrichment at regulatory regions. c Smoking-DMPs enrichment at chromatin states d TFBSs enriched in hypermethylated smoking-DMPs in more than 3 chromatin states. e Percentage of smoking-DEGs associated with at least one smoking-DMP, and vice versa. f Correlation of methylation residuals in cg25648203 with expression residuals of AHRR. g Enrichment of significantly correlated DMPs-DEGs in lung at regulatory regions. h Percentage of positive and negative correlations for hypomethylated enhancers (top), and the pathway in which these correlations are enrichment (bottom)

Fig. 6

Smoking and age DNA methylation effects. a Tissues with significant concordance in the direction of change (up- or down-regulation) for smoking-age-DMPs (chi-squared tests; FDR < 0.05). b Enrichment of smoking-age-DMPs at regulatory regions. c Enrichment of smoking-age-DMPs at chromatin states. d Shared TFBSs enriched in hypermethylation across more than 7 chromatin states for smoking-age-DMPs. TFs highlighted in bold are part of the Polycomb repressive complex

Fig. 7

Molecular and histological impact of smoking cessation. a Smoking-DEGs classification into reversible, partially reversible, or non-reversible genes per tissue. b Smoking-DMP classification into reversible, partially reversible, and non-reversible. c Examples of a reversible gene, d a partially reversible gene, and e a non-reversible gene. f Smoking-DEGs log fold change in never vs. ex-smokers and in ex-smokers vs. smokers. p-value obtained from a Wilcoxon test. g Smoking-DMP log fold change in never vs. ex-smokers and in ex-smokers vs. smokers. p-value obtained from a Wilcoxon test. h Classification of ex-smoker individuals into smokers or never smokers per tissue in gene expression, histology, and methylation. The green bars represent the number of re-classified ex-smoker samples. For gene expression, only highly accurate models (AUC > 0.85) were considered for the classification of ex-smokers. Asterisks correspond to FDR < 0.05 for binomial tests